R Notebook

To Do

• Why is CMS_20 empty?
• Double check that CMS_41 is yes/no

library(stringr)
library(tidyverse)
library(data.table)

ConnectCalls <- fread("/Users/guslipkin/Downloads/Florida Poly- Senior Capstone Group/ConnectCalls.csv", na.strings = "")
colnames(ConnectCalls) <- c("EncounterID",
                            "LastStatus",
                            "UpdatedOn",
                            "AQ1", "AQ1_Date",
                            "AQ2", "AQ2_Date",
                            "AQ3", "AQ3_Date",
                            "AQ4", "AQ4_Date",
                            "AQ5", "AQ5_Date",
                            "AQ6", "AQ6_Date",
                            "AQ7", "AQ7_Date",
                            "AQ8", "AQ8_Date",
                            "AQ9", "AQ9_Date",
                            "AQ10", "AQ10_Date",
                            "AQ11", "AQ11_Date",
                            "AQ12", "AQ12_Date",
                            "AQ13", "AQ13_Date",
                            "AQ14", "AQ14_Date",
                            "AQ15", "AQ15_Date",
                            "AQ16", "AQ16_Date",
                            "AQ17", "AQ17_Date",
                            "AQ18", "AQ18_Date",
                            "AQ19", "AQ19_Date",
                            "AQ20", "AQ20_Date",
                            "AQ21", "AQ21_Date",
                            "AQ22", "AQ22_Date",
                            "AQ23", "AQ23_Date",
                            "AQ24", "AQ24_Date")

factorCols <- c("EncounterID", "LastStatus", 
                names(ConnectCalls)[names(ConnectCalls) %like% "AQ[0-9]{1,2}$"])
ConnectCalls[, (factorCols) := lapply(.SD, as.factor), .SDcols = factorCols]

dateCols <- c("UpdatedOn", 
              names(ConnectCalls)[names(ConnectCalls) %like% "AQ[0-9]{1,2}_Date"])
dateFun <- function(x) as.Date(x, format = "%m/%d/%y")
ConnectCalls[, (dateCols) := lapply(.SD, dateFun), .SDcols = dateCols]

head(ConnectCalls)

aqCols <- c("EncounterID", "LastStatus", "UpdatedOn",
            names(ConnectCalls)[names(ConnectCalls) %like% "AQ[0-9]{1,2}$"])
ConnectLongerAQ <- melt.data.table(ConnectCalls[!is.na(EncounterID), ..aqCols],
                                 measure.vars = paste0("AQ", 1:24),
                                variable.name = "AQ_Number",
                                value.name = "AQ_Value")

aqCols <- c("EncounterID", "LastStatus", "UpdatedOn",
  names(ConnectCalls)[names(ConnectCalls) %like% "AQ[0-9]{1,2}_Date"])
ConnectLongerAQDate <- melt.data.table(ConnectCalls[!is.na(EncounterID), ..aqCols],
                                       measure.vars = paste0("AQ", 1:24, "_Date"),
                                       variable.name = "AQ_Number",
                                       value.name = "AQ_Date")

ConnectLongerAQDate[, AQ_Number := str_replace(AQ_Number, "_Date", "")]

ConnectLonger <- merge(ConnectLongerAQ, ConnectLongerAQDate)[!is.na(AQ_Value)]
rm(ConnectLongerAQ, ConnectLongerAQDate)
head(ConnectLonger)

ConnectLonger[, .N, by = "AQ_Value"][order(desc(N))] %>%
  ggplot(aes(x = AQ_Value, y = N, label = N)) +
  geom_col(fill = "lightblue") +
  geom_text(size = 5, position = position_stack(vjust = 0.5)) +
  theme(axis.text.x = element_text(angle = 90))

Surveys <- fread("/Users/guslipkin/Downloads/Florida Poly- Senior Capstone Group/Surveys.csv", na.strings = "")

factorCols <- c(names(Surveys)[! names(Surveys) %in% c("AdmitDate",
                                                       "CMS_Reportable",
                                                       "MortalityFlag")])
Surveys[, (factorCols) := lapply(.SD, as.factor), .SDcols = factorCols]

dateCols <- c("AdmitDate")
dateFun <- function(x) as.Date(x, format = "%m/%d/%y")
Surveys[, (dateCols) := lapply(.SD, dateFun), .SDcols = dateCols]

Surveys$CMS_Reportable <- ifelse(Surveys$CMS_Reportable == "Yes", TRUE, FALSE)
Surveys$MortalityFlag <- ifelse(Surveys$MortalityFlag == "Y", TRUE, FALSE)

logicalCols <- c("CMS_12", "CMS_18", "CMS_21")
logicalFun <- function(x) ifelse(x == 1, TRUE, FALSE)
Surveys[, (logicalCols) := lapply(.SD, logicalFun), .SDcols = logicalCols]
Surveys <- Surveys[, -c("CMS_42", "CMS_43", "CMS_44")]

head(Surveys)

cmsCols <- c("SurveyID", 
            names(Surveys)[names(Surveys) %like% "CMS_[0-9]{1,2}$"])
SurveysLonger <- melt.data.table(Surveys[!is.na(SurveyID), ..cmsCols],
                                 measure.vars = paste0("CMS_", c(1:3, 6:8, 10:13, 
                                                                 18:28, 30, 33:34, 
                                                                 37:41)),
                                 variable.name = "CMS_Number",
                                 value.name = "CMS_Response")[!is.na(CMS_Response)]

Warning in melt.data.table(Surveys[!is.na(SurveyID), ..cmsCols], measure.vars = paste0("CMS_",  :
  'measure.vars' [CMS_1, CMS_2, CMS_3, CMS_6, ...] are not all of the same type. By order of hierarchy, the molten data value column will be of type 'character'. All measure variables not of type 'character' will be coerced too. Check DETAILS in ?melt.data.table for more on coercion.

SurveysLonger[!(CMS_Number %in% c("CMS_12", "CMS_18", "CMS_21", "CMS_23"))] %>%
  .[, CMS_Response := as.factor(CMS_Response)] %>%
  .[, .N, by = list(CMS_Number, CMS_Response)] %>%
  ggplot(aes(x = CMS_Number, y = N, fill = factor(CMS_Response, levels = c(5:0)))) +
  geom_bar(position = "fill", stat = "identity") +
  theme(axis.text.x = element_text(angle = 90),
        legend.title = element_blank())

SurveysLonger[(CMS_Number %in% c("CMS_12", "CMS_18", "CMS_21"))] %>%
  .[, CMS_Response := as.factor(CMS_Response)] %>%
  .[, .N, by = list(CMS_Number, CMS_Response)] %>%
  ggplot(aes(x = CMS_Number, y = N, fill = CMS_Response)) +
  geom_bar(position = "fill", stat = "identity") +
  theme(axis.text.x = element_text(angle = 90),
        legend.title = element_blank())

SurveysLonger[CMS_Number == "CMS_23"] %>%
  .[, CMS_Response := as.factor(CMS_Response)] %>%
  .[, .N, by = list(CMS_Number, CMS_Response)] %>%
  ggplot(aes(x = CMS_Number, y = N, fill = factor(CMS_Response, levels = c(10:0)))) +
  geom_bar(position = "fill", stat = "identity") +
  theme(axis.text.x = element_text(angle = 90),
        legend.title = element_blank())

Readmissions <- fread("/Users/guslipkin/Downloads/Florida Poly- Senior Capstone Group/Readmissions.csv", na.strings = "")

colnames(Readmissions) <- str_replace_all(names(Readmissions), 
                                          pattern = " ", 
                                          replacement = "")

factorCols <- c(names(Readmissions)[! names(Readmissions) %in% c("IndexDischDate", 
                                                                 "ReadmitDate", 
                                                                 "DaysBetweenAdmssions", 
                                                                 "ReadmitDischargeDate", 
                                                                 "ReadmitInfo.Age")])
Readmissions[, (factorCols) := lapply(.SD, as.factor), .SDcols = factorCols]

dateCols <- c("IndexDischDate", "ReadmitDate", "ReadmitDischargeDate")
dateFun <- function(x) as.Date(x, format = "%m/%d/%y")
Readmissions[, (dateCols) := lapply(.SD, dateFun), .SDcols = dateCols]

head(Readmissions)

Encounters <- fread("/Users/guslipkin/Downloads/Florida Poly- Senior Capstone Group/Encounters.csv", na.strings = "")

colnames(Encounters) <- str_replace_all(names(Encounters), 
                                        pattern = " ", 
                                        replacement = "")

factorCols <- c(names(Encounters)[! names(Encounters) %in% c("AdmitDate", 
                                                             "DischargeDate", 
                                                             "Age")])
Encounters[, (factorCols) := lapply(.SD, as.factor), .SDcols = factorCols]

dateCols <- c("AdmitDate", "DischargeDate")
dateFun <- function(x) as.Date(x, format = "%m/%d/%y")
Encounters[, (dateCols) := lapply(.SD, dateFun), .SDcols = dateCols]

Encounters[EncounterID %in% c("40347"), Age := 43]
Encounters[!(EncounterID %in% c("22924", 
                                "34029",
                                "35544",
                                "44683",
                                "56159",
                                "68750")) | Age <= 110]

head(Encounters)

merge(Encounters, 
      merge(Surveys,
            merge(ConnectCalls,
                  Readmissions,
                  by = "EncounterID"),
            by = "EncounterID"),
      by = "EncounterID")